Systems and Methods for Predicting and Optimizing Physiological Performance

This application is a continuation of U.S. application Ser. No. 17/400,565 filed Aug. 12, 2021, which claims the benefit of U.S. application Ser. No. 16/105,569, filed Aug. 20, 2018 (pending), which claims the benefit of and priority to co-pending U.S. Provisional Application Ser. No. 62/547,449 filed Aug. 18, 2017. The contents of these applications are incorporated herein by reference in their entirety.

The invention described herein may be manufactured and used by or for the Government of the United States for all governmental purposes without the payment of any royalty.

The present invention relates generally to assessing individual and team physiology, physical activity, cognitive activity, and so forth. More particularly, the present invention relates to correlating individual and team physiology, physical activity, cognitive activity, and so forth to performance.

Countless physiological and activity (e.g., body movement) related parameters can now be measured non-invasively using various commercially-available wearable sensor and related software technologies including, but not limited to, products under the trade names Fitbit®, Apple Watch®, Zebra Motionworks®, Zephyr™ Performance Systems, Omegawave®, etc., Examples of the physiological parameters that are often tested include heart rate, heart rate variability (sympathetic and parasympathetic), respiration rate, blood oximetry, etc. Activity-related parameters include change of direction, acceleration, distance ran/walked, speed, explosiveness, etc. Additional parameters may be derived from physiological parameters, activity-related parameters, or combinations thereof with personal data regarding the user. For example, mostly all commercially-available fitness tracking products estimate caloric burn using height, weight, a level physical activity detected, and physiological parameters (such as a heart rate during the physical activity). Similarly, Omegawave® includes a “CNS” (central nervous system) score that uses “DC-Potential measurement to monitor and manage signs of fatigue in your Central Nervous System.” The Zephyr™ system uses “physiological load” (defined generally as a cumulative index of effort based on heart rate over a period of time), “physiological intensity” (defined generally as an instantaneous index of effort based on heart rate at that moment), “mechanical load” (defined generally as a cumulative index of effort based on acceleration over a period of time), “mechanical intensity” (defined generally as an instantaneous index of effort based on acceleration at a particular moment), “training load” (an average of the mechanical and physiological load), and “training intensity” (average of the mechanical and physiological intensity), to name a few.

While these commercially-available products are good at collecting and tracking many types of data for a particular individual, there remains a need for improvement as to how this data is used to predict and/or optimize performance. For example, the current commercially-available products are typically focused on displaying data being collected to give immediate feedback during training. In other words, the conventional products are typically focused on providing a “snap shot” of a particular individual's health at a particular moment in time without providing information that could be used to optimize the individual's future performance (over the next days or weeks, for example).

Another deficiency of conventional fitness tracking devices is that the comparing of each particular individual's data to so-called “normal ranges” that are based solely or primarily on physical and measurable information provided by the particular individual to the device (for example, age, height, weight). However, analysis of the data collected is quite user-dependent and “normal ranges” often specific to the particular individual. While certain evaluations are conventionally used to determine an individual's performance (such as physical battery tests, cognitive scores, blood biomarkers, heart rate, heart rate variability, autonomous and central nervous system responses, etc.), the results from each test is highly variable by individual. As a result, data obtained from these tests, when compared against normal ranges generalized to a large population of similar age, height, and weight, will not provide a complete picture for the particular individual.

Another difficulty with these conventional devices is the abundance of data, which complicates the sorting and identifying of which parameters are most important or significant for the particular individual. Moreover, parameters are activity dependent. That is, a parameter may be important for one type of sport or activity but not for a different sport or activity. What is needed therefore is a system for identifying parameters importance for a particular individual based on the sport or activity the individual is competing. Further, there is a need for methods of using the data collected for these important parameters to predict and optimize future performance for that same type of sport or activity.

Another issue with commercially-available fitness tracking products is that when a measured parameter is out of the normal range (e.g., heart rate is higher than normal), the devices does not providing information on how to bring the parameter back into the normal range apart from changing the current activity being performed (e.g., heart rate is high, so take a rest or lower activity level to lower the heart rate). Therefore, it would be further advantageous for the system to not only identify the important or significant parameters for a particular sport or activity, but to also identify certain manipulatable parameters (those parameters that may be changed directly or indirectly) that are best suited to keep the particular individual within an optimal range.

Still another deficiency of commercially-available fitness tracking products with respect to performance is that the data collected focuses on the particular individual's health without considering how the particular individual's health relates to a group or team. While it may be easy to adjust an individual's training schedule for individual activities (such as sports including tennis, running, golf, boxing, etc.), adjusting the training schedules for team sports are more difficult because the schedules are designed around the entire team being able to participate. For example, a football team often implements new offensive plays and formations in the week leading up to a particular game based on the particular style of defense of the upcoming opponent. Conventionally, the plays are practiced to “game speed” with the team's offensive competitors so as to simulate the upcoming game and better prepare the offense as a whole. On the other hand, if certain competitors are not performing optimally (are injured or sore from a previous game), then the team's offensive performance may be dictated more by a degree of soreness on game day as compared to the activity level of practice. In such instances, it may be advantageous to lower the practice speed or level to permit the team's offense to be rested for optimal performance on game day.

As a result, there remains a need for a system that may identify important or significant parameters based on a particular sport or activity, for individuals, teams, or both, using data collected from individuals of the team to design training schedules that optimizes performance of the individual, team, or both.

The present invention overcomes the foregoing problems and other shortcomings, drawbacks, and challenges of conventional fitness tracking products with respect to optimizing individual performance, team performance, or both. While the invention will be described in connection with certain embodiments, it will be understood that the invention is not limited to these embodiments. To the contrary, this invention includes all alternatives, modifications, and equivalents as may be included within the spirit and scope of the present invention.

According to one embodiment of the present invention, the above and other needs are met by a method for predicting physical performance for a future event comprising (a) selecting one or more performance parameters for the future event indicative of physical performance; (b) collecting data for the one or more performance parameters selected in step (a) from a plurality of past events that are substantially similar to the future event; (c) collecting data for a plurality of training parameters from prior to each of the plurality of past events; (d) for each of the one or more performance parameters selected in step (a), comparing the collected data for the plurality of training parameters of step (c) to the collected data for the performance parameter of step (b) for each of the plurality of past events to determine which of the plurality of training parameters are statistically significant training parameters; (e) determining an optimal performance range for one or more of the statistically significant training parameters; (f) collecting data for the one or more statistically significant training parameters having an optimal performance range as determined in step (e) prior to the future event; and (g) for each of the one or more statistically significant training parameters having an optimal performance range determined in step (e), predicting performance for the future event by comparing the collected data in step (f) with the optimal performance range determined in step (e).

According to certain embodiments, the method further includes optimizing performance for the future event by (h) determining which of the statistically significant training parameters are able to be manipulated; and (i) providing a training program for manipulating one or more of the statistically significant training parameters determined to be manipulatable in step (h) to conform to the optimal performance range of the readiness index developed in step (e) prior to the future event. In certain embodiments, step (h) includes determining whether at least one of the statistically significant training parameters is able to be manipulated indirectly by comparing the collected data of step (c) for the statistically significant training parameter with the collected data of step (c) for at least one of the other training parameters to determine whether the other training parameter is a statistically significant training variable for the statistically significant training parameter.

According to some embodiments, the one or more performance parameters selected in step (a) include one or more team-based metrics, step (b) includes collecting data for the one or more team-based metrics, and step (c) includes collecting individual-based training parameter data for each member of the team and converting the collected data for each of the training parameters to arrive at team-based training parameter data for comparison with the collected data for the team-based metrics of step (b). In certain embodiments, the converted team-based training parameter data of step (c) for each of the training parameters is compared in step (d) to the collected data for the team-based metrics of step (b) to determine which of the training parameters are statistically significant to the team-based metrics selected in step (a).

According to some embodiments, the one or more performance parameters selected in step (a) include one or more individual-based performance parameters, step (b) includes collecting data for the individual-based performance parameters, and step (c) includes collecting individual-based training parameter for comparison in step (d) with the collected data for the individual-based performance parameter data of step (b).

According to some embodiments, the data collected in step (c) is grouped into a plurality of time periods based on how long the data was collected prior to the corresponding past event and step (d) includes determining the statistically significant training parameters for each of the plurality of time periods.

According to some embodiments, the plurality of past events are performed by the same individual or team of individuals as the future event. In some embodiments, the plurality of past events are assigned a plurality of characteristics indicative of the past event and saved to a database, the method further comprising selecting the plurality of past events for comparison of the collected data in step (d) by matching expected characteristics of the future event to one or more of the characteristics assigned to the plurality of past events.

According to another embodiment of the present invention, a method for optimizing physical performance for a future event includes (a) selecting one or more performance parameters for the future event indicative of physical performance; (b) collecting data for the one or more performance parameters selected in step (a) from a plurality of past events that are substantially similar to the future event; (c) collecting a plurality of heart rate variability values from prior to each of the plurality of past events; (d) collecting data for one or more training parameters from prior to each of the plurality of past events; (e) determining an optimal performance range for heart rate variability prior to the future event by comparing the collected data for the plurality of heart rate variability values of step (c) to the collected data for the one or more performance parameters of step (b) for each of the plurality of past events; (f) determining which of the one or more training parameters are statistically significant to heart rate variability by comparing the collected data for the one or more training parameters of step (d) to the collected data from the plurality of heart rate variability values of step (c) for each of the past events; and (g) providing a training program for manipulating heart rate variability to conform to the optimal performance range determined in step (e) by manipulating one or more of the training parameters determined to be statistically significant in step (f) prior to the future event to optimize performance during the future event.

According to certain embodiments, the data collected in step (c) is grouped into a plurality of time periods based on how long the data was collected prior to the corresponding past event and step (e) includes determining the optical performance range for heart rate variability for each of the plurality of time periods.

According to certain embodiments, the plurality of past events are performed by the same individual or team of individuals as the future event. According to some embodiments, the plurality of past events are assigned a plurality of characteristics indicative of the past event and saved to a database, the method further comprising selecting the plurality of past events for comparison of the collected data in step (e) by matching expected characteristics of the future event to one or more of the characteristics assigned to the plurality of past events.

According to yet another embodiment of the present invention, a system for optimizing physical performance for a future event includes a user interface configured for selecting one or more performance parameters for the future event indicative of physical performance, selecting a plurality of training parameters that are potentially indicative of physical performance for the future event, uploading collected data for the selected performance parameters from a plurality of past events that are substantially similar to the future event, and uploading collected data for the selected training parameters from prior to each of the plurality of past events. The system further includes a computing program configured to compare the collected data for the selected training parameters to the collected data for the selected performance parameters to determine which of the selected training parameters are statistically significant training parameters for each of the selected performance parameters, determine which of the training parameters are able to be directly manipulated, comparing the collected data for the statistically significant training parameter from the plurality of past events with the collected data for the training parameters able to be directly manipulated to determine which of the directly manipulatable training parameters are a statistically significant training variable for the at least one statistically significant training parameter that is unable to be directly manipulated, and provide a training program for indirectly manipulating the statistically significant training parameter that is unable to be directly manipulated by directly manipulating one or more of the statistically significant training variables.

According to certain embodiments, the computing program is further operable to determine an optimal performance range for the at least one statistically significant training parameter that is determined to be unable to be directly manipulated, wherein the training program includes manipulating one or more of the statistically significant training variables such that the statistically significant training parameter conforms to the optimal performance range prior to the future event.

According to some embodiments, the data collected for the selected training parameters is grouped into a plurality of time periods based on how long the data was collected prior to the corresponding past event and the computing program is further operable to determine which of the selected training parameters are statistically significant training parameters for each of the plurality of time periods.

According to some embodiments, they system further includes a database for storing the collected data of the plurality of past events based on a plurality of characteristics indicative of the past event, the user interface further configured for selecting the plurality of past events for comparison of the collected data by matching expected characteristics of the future event to one or more of the characteristics of the plurality of past events.

According to certain embodiments, the system further includes one or more wearable fitness monitors for collecting data for one or more of the selected training parameters.

According to certain embodiments, the one or more performance parameters includes one or more team-based metrics and the plurality of training parameters include individual-based training parameters, the computing program further configured to convert the collected data for each of the individual-based training parameters to arrive at team-based training parameter data for comparison with collected data for the team-based metrics to determine the statistically significant training parameters for a team.

According to certain embodiments, the one or more performance parameters includes one or more individual-based performance parameters and the plurality of training parameters include individual-based training parameters for determining statistically significant training parameters for the individual by comparing collected data for the individual-based training parameters with collected data for the individual based performance parameters.

According to another embodiment of the present disclosure, a system for optimizing human physiological performance with respect to a set of physiological and activity-related parameters related to a particular event includes a display configured to operate as a user interface capable of receiving a user input and displaying an output to the user, a memory, storing one or more processor-readable instructions, and a processor, capable of receiving user inputs and displaying user outputs on the display based on the one or more processor-readable instructions. The instructions are configured to cause the system to: receive a selection of one or more individual performance parameters for a future instance of the particular event from the user interface, the individual performance parameters indicative of a desired physical performance of an individual for the particular event; receive a selection of one or more team performance parameters for a future instance of the particular event from the user interface, the team performance parameters indicative of a physical performance of a team for the particular event; collect data related to the individual performance parameters and the team performance parameters from past occurrences of the particular event; identify a subset of training parameters, from within a collection of training parameter data associated with past occurrences of the particular event or related events, that exhibit a statistically significant correlation with the selected individual performance parameters and the selected team performance parameters; and provide a training program, accessible on the user interface, listing each of the optimal training parameters and recommending a training routine for optimizing each of the individual performance parameters and the team performance parameters.

According to another embodiment of the present disclosure, a method for optimizing human physiological performance with respect to a set of physiological and activity-related parameters related to a particular event, the method comprising: receiving, via a user interface, a selection of one or more individual performance parameters for an instance of the particular event, the individual performance parameters indicative of a desired physical performance of an individual for the particular event; receiving, via the user interface, a selection of one or more team performance parameters for an instance of the particular event, the team performance parameters indicative of a physical performance of a team for the particular event; collecting data related to the individual performance parameters and the team performance parameters from past occurrences of the particular event; accessing training parameter data associated with the past occurrences of the particular event; identifying, from the training parameter data, a subset of training parameters that exhibit a statistically significant correlation with the individual performance parameters and the team performance parameters; generating a training program that lists each of the identified training parameters and recommends a training routine for optimizing the individual performance parameters and the team performance parameters; and providing the training program to a user via the user interface.

According to yet another embodiment of the present disclosure, a system for facilitating improvement of human physiological performance relative to a particular event, the system comprising: a display configured as a user interface for receiving user input and presenting information to a user; a memory storing one or more processor-executable instructions; and a processor operably coupled to the display and the memory. The processor is configured to execute the instructions to: obtain, via the user interface, input identifying one or more individual performance parameters associated with a future instance of the particular event, the individual performance parameters indicative of a desired outcome for an individual participant in the event; obtain, via the user interface, input identifying one or more team performance parameters associated with the future instance of the particular event, the team performance parameters indicative of a desired outcome for a group of participants in the event; access stored data representing historical instances of the particular event or related events, the historical data comprising training parameter data and corresponding performance outcomes; determine a set of training parameters, from the historical training parameter data, that exhibit a predictive relationship with the identified individual performance parameters and team performance parameters, based on a statistical analysis of the historical data; generate and display, via the user interface, a training program including training recommendations corresponding to the determined training parameters, the training recommendations being configured to guide the individual in achieving the identified individual performance parameters and team performance parameters.

Additional objects, advantages, and novel features of the invention will be set forth in part in the description which follows, and in part will become apparent to those skilled in the art upon examination of the following or may be learned by practice of the invention. The objects and advantages of the invention may be realized and attained by means of the instrumentalities and combinations particularly pointed out in the appended claims.

It should be understood that the appended drawings are not necessarily to scale, presenting a somewhat simplified representation of various features illustrative of the basic principles of the invention. The specific design features of the sequence of operations as disclosed herein, including, for example, specific dimensions, orientations, locations, and shapes of various illustrated components, will be determined in part by the particular intended application and use environment. Certain features of the illustrated embodiments have been enlarged or distorted relative to others to facilitate visualization and clear understanding. In particular, thin features may be thickened, for example, for clarity or illustration.

Referring to, a flowchart illustrating a processis depicted for predicting performance, optimizing performance, or both for a future event according to one embodiment of the present invention. For purposes of the present invention, an “event” is any type of physical and/or mental activity involving skill in which a level of performance may be objectively or subjectively quantified. The event may be related to an individual sport (such as tennis, golf, cycling, running, etc.) or a team sport (such as football, basketball, baseball, lacrosse, etc.). The event may be a singular game for the particular sport or a collection of games (such as a tournament). For example, a golf tournament may include four rounds of 18 holes, each round commenced and completed on a separate day. Each round or day may be considered a separate event or the tournament, as a whole, may be considered an event, depending on a desired analysis. An event may also be related to countless other types of physical and/or mental activities not typically considered to be a sport, such as a military activity (e.g., a military mission or operation), artistic performance (such as dance, musical and theatrical performances), or largely cognitive activities (such as a chess match or a standardized examination). In some embodiments, the event may be primarily a physical activity or considered more of a physical activity than a mental activity.

For purposes of the present invention, a “past event” refers to an event in which data is collected prior to the “future event.” In other words, the term “past” is used to refer to events occurring prior to the future event for which performance is being predicted, optimized, or both. In some embodiments, past events are performed by the same individuals that will perform in the future event. That is, for individual sports, the same athlete is assessed in both past and future events; for team sports, the same group of athletes is assessed in both past and future events. According to some embodiments, the past events may be substantially similar to the future event. For example, if the future event is a basketball game, then a substantially similar past event may be a previously played basketball game, a currently played basketball game, a scrimmage, a practice, or combination thereof. If the future event is a football game, then past events may include previous football game(s), preferably by the same team members that will be competing in the future event. If the future event is a tennis match, then the past events may include a previous tennis match or matches against an opponent that is the same or similar to the opponent that is competing in the future event.

According to stepof process, one or more performance parameters for the future event are selected. Selection of performance parameters may be based on ability to be indicative of or relatedness to physical performance for the future event. For example, if the future event is a singles tennis match, then suitable performance parameters may include individual metrics, which may include an ultimate outcome of the match (i.e., win vs. loss), personal statistics that are conventionally recorded kept during a tennis match and that are indicative of performance (such as a number of aces, a number of unforced errors, a number of sets won, etc.), or a self-evaluation (for example, feeling tired, sore, mental fog, headache, etc.). If the future event is a team event, such as a basketball game, then suitable performance parameters may include individual metrics (including those provided above), team-based metrics, or both. Exemplary team-based metrics for the basketball game may include an ultimate outcome of the game (i.e., win vs. loss), team statistics that are conventionally recorded during the basketball game and that are indicative of team performance (such as points scored, points scored by opponent, team fouls, team assists, team rebounds, etc.), or team evaluation (for example, degree of communication, team unity, etc.). Individual metrics for the basketball game may be similar to the team metrics, but kept at an individual level (such as the individual's points scored, assists, rebounds, etc.). For team-based events, evaluation of performance parameters comprising both team-based and individual metrics is preferably selected.

In addition to the selection of performance parameters based on the type of future event being predicted and/or optimized, training parameters are selected to be tracked in step. Selection of training parameters may be based, in part, on a belief that the selected training parameters may be correlated or related to the performance parameters selected in step. According to certain embodiments, selection of the training parameters may be based, at least in part, on the type of data acquired or tracked by the conventional fitness tracker available to or used by the individual or team. Exemplary training parameters may include physiological data (such as heart rate, heart rate variability, respiration rate, blood oximetry, etc.) and activity-related data (such as a number of direction changes, distance covered, average speed, explosiveness, etc.). For purpose of the present invention, activity-related data may also include behavioral-related data (such as an amount of sleep, number of meals, fluid intake, overall soreness, number of sore areas, etc.). Some embodiments may include training parameters, such as caloric intake, that may be derived from combinations of multiple physiological parameters and/or activity-related parameters.

Referring to stepsand, data is collected or otherwise acquired with respect to the selected performance parameters of stepand the selected training parameters of stepfor past events. For the performance parameters, data may include individual metrics, team metrics, or both, which are described above. For the training parameters, data may typically be individual-based data, such as those described above. When individual-based data for the training parameters is being correlated to team-based metrics, then a mean, a median, a weighted-average, or other statistical sampling of the individual-based data may be used for the particular training parameter. For example, if the team-based metric is whether the team won or lost and the training parameter is an amount of sleep, then the individual-based data for each team member (or each team member that participated in the past events or the future event) may be converted to team-based data (e.g., summed or averaged) for the training parameter to be correlated with the team-based data for the performance parameter for the past event.

It would be readily appreciated by those of ordinary skill in the art having the benefit of the disclosure made herein that the more related past events are to the future event, the better the predictive quality may be for the system (so long as a sufficient sample size of past events has been collected). Thus, according to certain embodiments, each past event may be assigned a characteristic indicative of the past event and loaded to a database of past events with the characteristics being identified for each of the past events. Once a significant sample size of data for past events is loaded to the database with the characteristic, a user may select certain past events from the database for correlation to a future event by narrowing the past events to those having a characteristic similar to the future event. For example, if the future event is a basketball game against a high-quality opponent as compared to a weaker opponent, then a user may select other past events (or previous games) by the same team against similar high-quality opponents to be used as past events instead of all evaluating against all past events (or all basketball games that the team has competed in). Characteristics that may be used to organize and select past events for correlation to the future event are numerous and may include, for example, the time of day in which event takes place, home game versus road game, time of travel to event, proximity in time to other events, weather during event (e.g., temperature, rain, sun, snow, humidity, etc.), expected duration of the event, and so forth.

For team sports, past events may be characterized by the players involved during the event. For example, if a star player was injured or otherwise not available during a past event, then that past event may be removed from the analysis if the star player is expected to participate in the future event. Similarly, a new individual to a team may be initially correlated with past events in which the new individual did not participate, and then data collected during the future event in which the new individual did participate may be loaded to the database for validating the initial correlation.

It would be readily understood by the skilled artisan that once the future event has concluded, data collected with respect to the future event may be used as a “past event.” In this way, the predictive quality of the data and the system may be continually updated and improved.

With respect to stepand data being collected for the selected performance parameters for past events, the data may be either objectively quantified or subjectively quantified. For example, if the event is a basketball game, team and/or individual performance parameters such as win vs. loss, number of assists, number of points, points by opponent, etc. are able to be objectively quantified by collecting the data during the past event. However, particularly for individual sports or individual data within a team sport, subjective quantification may be used such as “rate your performance today on a scale of 1-10” using a subject survey or third-party ratings (e.g., coach ratings) after the past event is completed.

Similarly, with respect to stepand data being collected for the selected training parameters for past events, the data is preferably objectively quantified when possible and subjectively quantified when needed. Most of the training parameter data should be able to be objectively quantified using commercially-available fitness trackers. On the other hand, other training parameter data such as number of meals eaten and fluid intake may need to be objectively quantified data using a subject survey. The subject survey may also be used to collect the data for subjectively quantified training parameter data such as total soreness (e.g., “rate how sore you are on a scale of 1-10”), number of sore areas, emotional stress, etc.

With continuing reference to step, and according to some embodiments of the present invention, training parameter data may be grouped into a plurality of time periods based on when the data was collected as compared to the past event to which the training parameter data corresponds. For example, training parameter data (such as change of direction, max acceleration, distance ran, etc.) may be collected during training sessions for a few days prior to the future event, which may then be grouped by the number of days in which the training session was held before the corresponding past event (e.g., one data point for the number of times the individual changed directions during a training session four days before the event, another data point for the number of times the athlete changed directions during a training session three days before the event, and so forth). Data for certain training parameters, particularly physiological parameters (such as heart rate and heart rate variability), may also be collected a plurality of times during a day (e.g., one data point for heart rate when the athlete woke up on the morning four days before the event, one data point for heart rate before the athlete went to bed four nights before the event, . . . , and one data point for heart rate on the morning of the event).

Once training parameter data and corresponding performance parameter data are collected in stepsandfor a sufficient number of past events, stepof the processdetermines which of the selected training parameters of stepare statistically significant to the selected performance parameters of step. When data for training parameters is grouped into a plurality of time periods, stepmay further include determining the statistically significance of training parameters for each of the plurality of time periods. For purposes of the present invention, one parameter is “statistically significant” to another parameter when the parameters pass a statistically significant test as known in the art to a certain predetermined confidence level. For example, according to certain embodiments, if a correlation between particular training parameter and a performance parameter has a p-value of 10% or less, then the training parameter may be considered to be statistically significant. In other words, in situations in which there is 10% or less chance that the relationship between the training parameter and the performance parameter was the result of chance, the training parameter may be considered to be statistically significant to the performance parameter. On the other hand, if the p-value is greater than 10%, then the parameter may be considered to be statistically insignificant. As data is collected for more and more past events, the predetermined confidence level (for example, the p-value) may be changed to require a greater confidence level (such as a p-value of 5% or less). In addition to p-values, the statistically significance may be determined by other statistics tests and techniques such as genetic algorithms, lass net models, etc.

The statistically significant training parameters identified in stepmay be used in a variety of ways in relationship to performance at future events. For example, referring to steps-, performance for a future event may be predicted to modify strategy for the future event. In this regard, in step, an optimal performance range for a training parameter may be determined for by comparing data associated with the particular training parameter for each past event to the performance parameter data of the same past event. Again, when data for the training parameter is grouped into a plurality of different time periods, then stepmay further include determining an optimal performance range for each time period of the plurality. The optimal performance range may be preferably determined by a statistical distribution type analysis (such as shown in a bell curve, z-table, etc.). Alternately, a prediction model may be used where potential values for the data are used in the models to determine how the performance parameter would change in response to the training parameter data. An optimal performance range may then be chosen based on the prediction model results. In step, data is collected for the particular training parameter prior to the future event. Then, in step, performance may be predicted by comparing the training parameter data from stepto the optimal performance range for the parameter determined in step.

As would be understood, the ability to predict performance may result in significant advantages. For example, as further explained below, it has been determined that heart rate variability is often a statistically significant training parameter when compared to performance parameters of different types of events. Thus, assuming heart rate variability is found to be statistically significant training parameter for an individual basketball player or a basketball team in stepand an optimal performance range for heart rate variability for the individual or team is determined in step, then a coach may decide to sit or otherwise adjust playing times for the individual basketball player during the future event if data collected during stepshows the individual basketball player's heart rate variability to be outside of the optimal performance range. In other words, the coach is able to recognize, ahead of time, when there is a statistically good chance that a player will have a less than optimal performance during a particular future event. Accordingly, the coach may use this information to determine whether the player should receive more or less playing time than normal during the game.

Similarly, referring to step, the optimal performance ranges determined in stepand training parameter data collected in stepmay also be used to optimize performance during a future event by detecting that a player or team is outside the particular optimal performance range that was determined in step. And so, continuing with the heart rate variability example above, assuming in stepthat heart rate variability was determined to be statistically significant two days before a game, then in step, it is then determined that a player's heart rate variability is outside the individual's and/or team's optimal performance range three-four days before the game, then in step, expected performance may be improved by proposing a training schedule that is designed to get the athlete's heart rate variability to conform to the optimal performance range.

Referring now to, a flowchart illustrating a processfor optimizing performance according to another embodiment of the present invention is described. Similar to the processof, processmay include determining which training parameters are statistically significant to the performance parameters for a particular type of event by selecting performance parameters in step, selecting training parameters in step, collecting performance parameter data for past events in step, collecting training parameter data for past events in step, and determining which training parameters are statistically significant to the performance parameters in step. To optimize performance, processfurther includes determining which of the statistically significant training parameters are manipulatable in step. Directly manipulatable parameters may include, for example, a number of reps (such as a number of changes in direction during a training session); indirectly manipulatable parameters may include, for example, results from level of training (e.g., overall soreness of an athlete may be found to be indirectly changed by adjusting a number of changes in directing during a training session). Thus, stepmay include determining which statistically significant training parameters may be changed directly and which statistically significant training parameters must be changed indirectly or are otherwise difficult to change directly.

According to certain embodiments, determining the manipulatable parameters in stepincludes comparing the statistically significant training parameters to the other training parameters to determine whether and which of the other training parameters may be statistically significant to a particular statistically significant training parameter. For example, supposing heart rate variability is found to be a statistically significant training parameter to performance in step, then while heart rate variability may difficult to manipulate directly, stepenables the identifying of training parameters that are best suited to manipulate the heart rate variability indirectly based on relatedness by statistical significance. That is, by determining which training parameters are statistically significant to heart rate variability according to a pre-determined confidence level. Accordingly, stepprovides the ability to determine which training parameters may be directly manipulated, that are not necessarily statistically significant to performance parameters, but that may be used to indirectly manipulate the training parameter that is unable to be manipulated directly.

As should be understood, certain training parameters could be found to be both a statistically significant training parameter for a particular performance parameter as well as a statistically significant training variable to another statistically significant training parameter.